Analysis of Momentum in Tennis Matches Based on Ensemble Learning

Huanyu Gong; Xuanxuan Guo; Zhuohui Gao

doi:10.62051/024qdq85

Authors

Huanyu Gong
Xuanxuan Guo
Zhuohui Gao

DOI:

https://doi.org/10.62051/024qdq85

Keywords:

Momentum; Random Forest; CVI; XGBoost.

Abstract

In tennis, player performance is often attributed to "momentum," pivotal for game control. We analyzed tennis rules, extracting performance indicators such as point differentials, dynamic scoring, and serving advantages. Utilizing a Random Forest model with player performance as the dependent variable, we mapped features to momentum and determined feature importance, achieving an AUC of 0.8796. We then introduced the Comprehensive Ability Volatility Index (CVI) based on factors like serve win rate and breakpoint success rate, identifying match-turning points. To capture these turning points, we integrated momentum, player fitness, and overall strength into an XGBoost model, which yielded an AUC of 0.8716. The model's feature importance output guided concise recommendations for strategic decision-making. Our research offers comprehensive strategies for athletes and coaches to consider momentum, physical fitness, and other match factors, enhancing guidance for daily training and match adjustments.

Downloads

Download data is not yet available.

References

[1] Depken C A, Gandar J M, Shapiro D A. Set-level strategic and psychological momentum in best-of-three-set professional tennis matches[J]. Journal of Sports Economics, 2022, 23 (5): 598 - 623.

[2] Morgulev E, Avugos S. Beyond heuristics, biases and misperceptions: the biological foundations of momentum (hot hand) [J]. International Review of Sport and Exercise Psychology, 2023, 16 (1): 155 - 175.

[3] Meier P, Flepp R, Rüdisser M, et al. Investigating the conditions for psychological momentum in the field: Evidence from men’s professional tennis [J]. Available at SSRN 3438123, 2019.

[4] Bühren C, Steinberg P J. The impact of psychological traits on performance in sequential tournaments: Evidence from a tennis field experiment [J]. Journal of Economic Psychology, 2019, 72: 12 - 29.

[5] Ötting M, Langrock R, Maruotti A. A copula-based multivariate hidden Markov model for modelling momentum in football [J]. AStA Advances in Statistical Analysis, 2021: 1 - 19.

[6] Sheykhmousa M, Mahdianpari M, Ghanbari H, et al. Support vector machine versus random forest for remote sensing image classification: A meta-analysis and systematic review [J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020, 13: 6308-6325.

[7] Sagi O, Rokach L. Approximating XGBoost with an interpretable decision tree [J]. Information sciences, 2021, 572: 522 - 542.

[8] Janssens A C J W, Martens F K. Reflection on modern methods: Revisiting the area under the ROC Curve [J]. International journal of epidemiology, 2020, 49 (4): 1397 - 1403.

[9] Muschelli III J. ROC and AUC with a binary predictor: a potentially misleading metric [J]. Journal of classification, 2020, 37 (3): 696 - 708.

[10] Zhang Y, Liao Q V, Bellamy R K E. Effect of confidence and explanation on accuracy and trust calibration in AI-assisted decision-making [C]//Proceedings of the 2020 conference on fairness, accountability, and transparency. 2020: 295 - 305.